As you may have read in recent posts about our project, CAMERA, we have been focusing on using artificial intelligence to analyse the mobility research landscape in Europe and to identify mobility research trends, gaps and needs. The CAMERA team has aimed to provide frameworks, insights and advice that may help shape the future of mobility and aviation research in Europe; in turn, the idea is that this may nudge Europe to reach its mobility goals, many of them outlined in the well-known Flightpath2050 document. We have been publishing our findings in annual reports—called Mobility Reports—and three of them can be found on CAMERA’s webpage.
On June 17, as part of Agency Research Team (ART) workshop on passenger-centred mobility (14 – 17 June), CAMERA presented its progress in applying artificial intelligence to the analysis of mobility research documents. In particular, we focused on mobility indicators across 11 Key Performance Areas (KPAs, heavily inspired by 11 KPAs defined by ICAO) as a way of capturing relevant mobility goals. We also focused on the techniques used to assess FP7 and H2020 research initiatives against those KPAs.
Faced with the challenge of having to analyse a large database of textual documents containing information on over 40,000 research initiatives in Europe, we opted to use information retrieval techniques to assess their relevance for mobility research goals and defined KPAs. Information retrieval is an area that involves the use of principles and algorithms to find relevant information in a collection of unstructured (usually textual) data; it proved to be very useful in the context of CAMERA. We developed several NLP-based models that helped us filter relevant mobility research initiatives, assess them against the defined mobility goals, and identify common research topics and trends. This process has also been automated up to a certain degree, which sped up our analysis of mobility research landscape in Europe in the past fifteen years and which enabled us to deliver a number of interesting insights and infographics.
In the workshop, after we presented on KPAs and information retrieval techniques, we asked participants to get their hands dirty by exploring some of the CAMERA results themselves and giving us feedback. As our information retrieval models are of probabilistic nature, and thus deliver results of probabilistic nature (i.e. approximate matches), introducing human experts and their feedback into the loop is extremely helpful. Moreover, through a series of stimulating open discussions, we discussed relevant mobility indicators for five KPAs (flexibility, predictability, digitalisation, environment and safety). Through those discussions, some interesting new mobility indicators appeared that were not included in the original framework of CAMERA—we are very keen to include them in the final version.
In order to facilitate analysis and foster discussion, we presented participants with the dashboard of various data visualisations within CAMERA. This dashboard currently allows users to explore CAMERA’s selected mobility research initiatives, scope them by their research topics and relevant KPAs, analyse received EC contribution, consortium size, geographical scope, and test many other functionalities. We invite you take a look and give us your feedback via the open form at the bottom of the page. We encourage any feedback (even “I didn’t like this colour” comments for new infographic ideas). We will aim to incorporate as much of it as possible into the final version of the dashboard. You can access it following this link.
If you were unable to attend the workshop, take a sneak peek at all the presentations here (CAMERA was on Day 4 of the workshop). Moreover, all the findings and insights, a ton of new interesting data visualisations, and the final dashboard are going to be presented in Mobility Report 4 (final one!) that will be published in October 2021. Hence, this is your last chance to participate and shape the scope of CAMERA with your input!
